diff --git a/fleetrec/core/factory.py b/fleetrec/core/factory.py
index 292a2e4c3ab5e0ccc1c757f0bb5a148b8f3d4eac..cfe7341d503657ed3e8926039c2660d907bd6979 100644
--- a/fleetrec/core/factory.py
+++ b/fleetrec/core/factory.py
@@ -33,15 +33,15 @@ class TrainerFactory(object):
     def _build_trainer(config, yaml_path):
         print(envs.pretty_print_envs(envs.get_global_envs()))
 
-        train_mode = envs.get_global_env("train.trainer")
+        train_mode = envs.get_runtime_envion("train.trainer")
 
         if train_mode == "SingleTraining":
             trainer = SingleTrainer(yaml_path)
         elif train_mode == "ClusterTraining":
             trainer = ClusterTrainer(yaml_path)
-        elif train_mode == "CtrTrainer":
+        elif train_mode == "CtrTraining":
             trainer = CtrPaddleTrainer(config)
-        elif train_mode == "UserDefineTrainer":
+        elif train_mode == "UserDefineTraining":
             train_location = envs.get_global_env("train.location")
             train_dirname = os.path.dirname(train_location)
             base_name = os.path.splitext(os.path.basename(train_location))[0]
diff --git a/fleetrec/core/utils/envs.py b/fleetrec/core/utils/envs.py
index eb866fba5b549f924e34d9bbe153fa9d3b381d7d..53d68ce4c43be2cc468a9118d0ffdcf498a761a2 100644
--- a/fleetrec/core/utils/envs.py
+++ b/fleetrec/core/utils/envs.py
@@ -12,12 +12,23 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-
+import os
 import copy
 
 global_envs = {}
 
 
+def set_runtime_envions(envs):
+    assert isinstance(envs, dict)
+
+    for k, v in envs.items():
+        os.environ[k] = v
+
+
+def get_runtime_envion(key):
+    return os.getenv(key, None)
+
+
 def set_global_envs(envs):
     assert isinstance(envs, dict)
 
@@ -87,4 +98,3 @@ def lazy_instance(package, class_name):
     model_package = __import__(package, globals(), locals(), package.split("."))
     instance = getattr(model_package, class_name)
     return instance
-
diff --git a/fleetrec/run.py b/fleetrec/run.py
index 1a79ce76fcaabaa0b5124d0ea5d58a420d94abd8..51dbd1fd998f169cc30607c30146da7da2801dcb 100644
--- a/fleetrec/run.py
+++ b/fleetrec/run.py
@@ -11,11 +11,10 @@ def run(model_yaml):
     trainer.run()
 
 
-def single_engine(model_yaml):
-    single_envs = {}
-    single_envs["singleTraning"] = True
-
+def single_engine(single_envs, model_yaml):
     print(envs.pretty_print_envs(single_envs, ("Single Envs", "Value")))
+
+    envs.set_runtime_envions(single_envs)
     run(model_yaml)
 
 
@@ -47,25 +46,30 @@ if __name__ == "__main__":
 
     if args.engine == "Single":
         print("use SingleTraining to run model: {}".format(args.model))
-        single_engine(args.model)
+        single_envs = {}
+        single_envs["train.trainer"] = "SingleTraining"
+
+        single_engine(single_envs, args.model)
     elif args.engine == "LocalCluster":
         print("use 1X1 ClusterTraining at localhost to run model: {}".format(args.model))
 
         cluster_envs = {}
-        cluster_envs["server_num"] = 1
-        cluster_envs["worker_num"] = 1
-        cluster_envs["start_port"] = 36001
-        cluster_envs["log_dir"] = "logs"
+        cluster_envs["train.server_num"] = 1
+        cluster_envs["train.worker_num"] = 1
+        cluster_envs["train.start_port"] = 36001
+        cluster_envs["train.log_dir"] = "logs"
+        cluster_envs["train.trainer"] = "SingleTraining"
 
         local_cluster_engine(cluster_envs, args.model)
     elif args.engine == "LocalMPI":
         print("use 1X1 MPI ClusterTraining at localhost to run model: {}".format(args.model))
 
         cluster_envs = {}
-        cluster_envs["server_num"] = 1
-        cluster_envs["worker_num"] = 1
-        cluster_envs["start_port"] = 36001
-        cluster_envs["log_dir"] = "logs"
+        cluster_envs["train.server_num"] = 1
+        cluster_envs["train.worker_num"] = 1
+        cluster_envs["train.start_port"] = 36001
+        cluster_envs["train.log_dir"] = "logs"
+        cluster_envs["train.trainer"] = "CtrTraining"
 
         local_mpi_engine(cluster_envs, args.model)
     else: